Inactin, L-glutamate, and bicuculline methiodide were obtained from Sigma Chemical (St. Louis, MO). Drugs used for microinjection were dissolved in artificial cerebrospinal fluid (aCSF) in which pH had been adjusted to 7.3–7.5 using sodium hydroxide or hydrochloric acid.

All surgical and experimental procedures were approved by the Institutional Animal Care and Use Committee of Wayne State University and conducted in accordance with the American Physiological Society's “Guiding Principles in the Care and Use of Animals.” All animals received food (Purina Lab Diet #5001, Purina Mills, Richmond, IN) and tap water ad libitum.

Forty-nine male Sprague–Dawley rats (initially 75–100 g, Harlan, Indianapolis, IN) were used for these studies. Animals were housed individually in cages with (physically active group, n = 24) or without running wheels (sedentary group, n = 25) for 8–12 weeks. Running wheels were purchased from a national vendor (Tecniplast, Eaton, PA; wheel diameter 34 cm diameter). Running distances were measured via bicycle computers (Sigma Sport, Olney, IL) calibrated to the diameter of the running wheel and recorded daily by laboratory personnel. Previous studies have demonstrated that rats allowed to run spontaneously on running wheels exhibit increases in maximal oxygen consumption, increases in heart weight-to-body weight ratio, and other exercise-related adaptations (Nelson et al., 2010).

Following 8–12 weeks of sedentary or physically active conditions, animals were instrumented acutely for brainstem microinjections while recording mean arterial pressure (MAP), HR, and LSNA (Mueller, 2007; Mischel and Mueller, 2011). Briefly, animals were anesthetized with isoflurane (2% in 100% O2) so that arterial and venous femoral catheters could be implanted to measure arterial pressure and for the administration of drugs, respectively. Following a midline abdominal incision, the abdominal aorta was gently retracted and a section of the lumbar chain was isolated caudal to the renal vein. The nerve was carefully placed on silver wire electrodes before encasing in polyvinylsiloxane gel (Darby Dental Supply Co., Inc., Westbury, NY). The abdominal incision was sutured around the electrode as it exited the peritoneal cavity. A ground wire was attached to the incision to reduce noise. A tracheostomy was performed and the animals were ventilated artificially with a mixture of isoflurane and oxygen (2% in 100% O2) via an endotracheal tube. Rats were placed in a prone position in a stereotaxic device (Kopf, Tujunga, CA) and following a midline incision at the level of the occipital bone, the brainstem was exposed by retracting the underlying muscles, performing a partial occipital craniotomy, and removing the atlanto-occipital membrane.

Once all surgical procedures were complete, an initial infusion of Inactin (thiobutabarbital sodium, 0.025 ml/min, 100 mg/kg, i.v.) was performed over 20–30 min during which the isoflurane was slowly reduced. Following the infusion, Inactin was given in supplemental doses (5 mg, i.v.) until an appropriate level of anesthesia was maintained in the absence of isoflurane. Anesthesia was deemed appropriate by lack of a withdrawal reflex to firm pinch. Animals continued to be ventilated (60–80 breaths/min) with a mixture of 100% O₂ and room air. Arterial blood gases were determined and were maintained in the physiological range (P_O2 > 100 mmHg, P_CO2 between 35 and 40 mmHg) by adjusting the rate or volume of the ventilator. Body temperature was maintained near 37°C via the use of a heating pad and measured throughout the experiment with a rectal thermometer. In order to diminish outside electrical noise, experiments were performed inside a Faraday cage.

We performed microinjections based on previously published techniques in our laboratory (Mueller and Hasser, 2006; Mueller, 2007). Briefly, once the animal was in the stereotaxic frame, the head was positioned at a downward angle such that calamus scriptorius was 2.4 mm caudal to the interaural line and the brainstem oriented in the horizontal plane (Kiely and Gordon, 1994; Moffitt et al., 2002). Triple-barrel glass micropipettes were drawn by a pipette puller to an outside tip diameter of 30–60 μm, filled with appropriate drugs, and inserted into the dorsal surface of the brainstem with the use of a surgical dissecting microscope. The RVLM was located with the following coordinates using calamus scriptorius as a reference point: 0.9–1.1 mm rostral and 1.7–2.2 mm lateral to calamus scriptorius, and 3.5–3.7 mm ventral to the dorsal surface of the medulla. Micropipettes were connected via polyethylene tubing to a commercially available picoejection system (Toohey Company, Fairfield, NJ). The volume of microinjections was monitored directly by visualizing the meniscus in each barrel with the aid of a compound microscope (150×) that contained a calibrated reticule. The RVLM was functionally identified by observing pressor and sympathoexcitatory responses to a standard dose and volume of glutamate (10 mM, 30 nl). Microinjection sites were marked with 2% Chicago sky blue dye (30 nl). At the end of the experiments, animals were overdosed with Fatal-Plus euthanasia solution (Vortech, Dearborn, MI, 0.2 ml) and brains were removed and placed in 4% phosphate buffered formalin solution. Following post fixation, brains were transferred to 30% sucrose for a minimum 48-h infiltration. The hindbrain was frozen and cut into 30 μm sections on a cryostat. Coronal sections were mounted on gel-coated slides and a bright-field microscope was used to determine the center of the dye spot and its location in the brainstem with the aid of a rat brain atlas (Paxinos and Watson, 2007). The dye spot location was represented graphically on a modified diagram from the rat atlas (Paxinos and Watson, 2007).

We tested the hypothesis that responses to excitation of the RVLM were enhanced in sedentary animals compared to animals that ran spontaneously on running wheels. Unilateral microinjections of specific concentrations of glutamate (1, 10, and 100 mM) were performed at a constant volume of 30 nl (30, 300, and 3000 pmol total, respectively). Specific concentrations were ejected from individual barrels of the triple-barrel micropipette, performed in a random order, and a minimum of 5 min of recovery time was allowed between responses. Control injections of vehicle (i.e., aCSF) produced little or no response in MAP, HR, and LSNA (see Results).

We hypothesized that responses to glutamate would be enhanced in sedentary rats in the absence of tonic GABAergic tone compared to animals that ran spontaneously on running wheels. To test this hypothesis, GABA receptors were blocked unilaterally with the GABA_A receptor antagonist, bicuculline (5 mM, 60 nl, or 300 pmol total) prior to subsequent injections of glutamate. Responses to glutamate were compared before and 5, 15, 30, and 45 min after bicuculline. The concentration of bicuculline used was based on previous studies in which bicuculline was shown to inhibit GABA_A receptor activation in the RVLM (Miyawaki et al., 2002; Moffitt et al., 2002; Horiuchi et al., 2004) or block arterial baroreflex-mediated changes in sympathetic nerve activity (Sun and Guyenet, 1985; Blessing, 1988; Dampney et al., 1988; Guyenet and Stornetta, 2004; Mueller, 2007).

Microinjections of glutamate were tested in the presence and absence of aCSF (60 nl) to determine the influence of vehicle, volume, or time on responses to repeat microinjections of glutamate. The time course and size of injections were similar to those used for bicuculline in Protocol #2.

A computer data acquisition system (Power Lab, ADInstruments, Colorado Springs, CO) was used to collect all experimental data. Raw LSNA was monitored on a Tektronix oscilloscope and a Grass preamplifier (P511) was used to amplify (20,000×) and filter LSNA (3 kHz low pass and 30 Hz high pass filter). LSNA was electronically rectified, integrated, and averaged using a time constant of 28 ms. Changes in LSNA were calculated as a percentage of control prior to each microinjection. The average amplified voltage (20,000×) was also compared between all animals as an estimate of baseline sympathetic nerve activity. Hexamethonium (30 mg/kg) and atropine methyl bromide (1 mg/kg, i.v.) were administered as ganglionic blocking agents to determine background noise. Background noise was subtracted from each animal and the remaining signal was defined as LSNA.

Baseline hemodynamic variables, body weights, and organ weights were analyzed by Student's t-test. MAP, HR, or LSNA changes to specific concentrations of glutamate were analyzed by Two-Way analysis of variance (ANOVA) with repeated measures. When ANOVA indicated a significant interaction, differences between individual means were assessed by post-hoc Holm-Sidak test according to a commercially available software package (SigmaStat 3.0, SPSS Inc., Chicago, IL). In one instance square root transformations were performed on MAP/Glutamate dose response data to achieve normality before Two-Way ANOVA. The transformation did not change the statistical outcome of the test (p > 0.05).

For protocols 2 and 3 (bicuculline and aCSF, respectively), data were combined initially and subjected to multivariate and repeated measures analyses using the General Linear Model (GLM) program in IBM SPSS for Windows (Version 19.0 Armonk, NY: IBM Corp). Based on both theory and the literature regarding the time course of bicuculline effects, our expectation was that any significant effects over the time points measured in both protocols (control, 5′, 15′, 30′) would be represented by the quadratic trend component (i.e., the two middle values would be higher than the first and fourth). Although preliminary analyses using both the multivariate and the repeated measures options gave similar results, the quadratic effects accounted for almost all of the between time variance. In the repeated measures analysis for the quadratic scores, we found both a significant time by experiment interaction and a time by group interaction (p < 0.05 for each). The former interaction justified testing the individual protocols (Bicuculline and aCSF) separately. We then performed simple main effects t-tests on the quadratic scores. As expected and borne out by the analysis, all of the effects revealed by ANOVA were due to bicuculline. As follow up to the significant group by quadratic time interaction (for bicuculline only), we analyzed the simple main effects of group at each time point using Holm–Sidak post-hoc tests provided by the Two-Way RM ANOVA. The results of these analyses are provided in the graphs and justify separation of the two protocols graphically.

For all analyses, a probability of p < 0.05 was considered statistically significant and p-values less than or equal to 0.1 for main effects are reported for clarity (Curran-Everett and Benos, 2004). Data are expressed as mean ± SEM.

Table 1 contains baseline characteristics in sedentary versus physically active animals. Physically active animals ran for an average of 10.4 ± 0.3 weeks which resulted in total distances ran of over 200 km (Table 1). Sedentary animals remained in their cages for similar periods of time without access to running wheels. Body weight was higher in sedentary animals (p < 0.05); however, there were no significant differences in baseline MAP, HR, or the average amplified LSNA voltage between groups. Similarly, baseline MAP, HR, and average amplified LSNA voltage was not significantly different between groups prior to each protocol (Table 2).

Representative MAP and LSNA responses in one sedentary and one physically active rat to RVLM microinjection of glutamate (30 nl, 10 mM) are shown in Figures 2A,C. Averaged peak changes in MAP, HR, and LSNA to microinjections of glutamate (30 nl, 1–100 mM) are shown in Figure 1. As in our previous studies (Mueller, 2007; Mischel and Mueller, 2011), glutamate produced concentration-dependent increases in MAP, HR, and LSNA (p < 0.001, main effect of dose for all three variables). The pressor responses to increasing concentrations of glutamate were not statistically different between sedentary and physically active animals (p = 0.10 for main effect) and the lack of a significant interaction did not allow for testing for differences at individual concentrations between groups (p = 0.872). HR responses to glutamate microinjections into the RVLM were small in general (<20 bpm) and there was neither a main effect of sedentary condition (p = 0.215) nor a significant interaction (p = 0.489) that would allow for testing differences at individual doses. Lastly, unlike our previous studies in which we demonstrated enhanced sympathoexcitation in splanchnic (Mischel and Mueller, 2011) or lumbar sympathetic nerves (Mueller, 2007) of sedentary animals, increases in LSNA were not significantly different between sedentary and physically active animals (p = 0.576 for main effect; p = 0.646 for interaction).

Representative MAP and LSNA responses in one sedentary and one physically active rat to RVLM microinjection of glutamate (30 nl, 10 mM) before and after bicuculline are shown in Figure 2. Averaged peak changes in MAP, HR, and LSNA to microinjections of glutamate (30 nl, 10 mM) with and without bicuculline are shown in Figure 3. As in Protocol #1, control responses to glutamate (30 nl, 10 mM) produced increases in MAP and LSNA and small changes in HR (<10 bpm) in both sedentary and physically active animals. Unilateral injections of bicuculline increased baseline MAP (Sedentary Δ9 ± 2 mmHg; Physically Active Δ13 ± 2 mmHg), HR (Sedentary –Δ7 ± 6 bpm; Physically Active Δ5 ± 2 bpm), and LSNA (Sedentary Δ 3 ± 4%; Physically Active Δ9 ± 3%) prior to the subsequent injection of glutamate at 5 min. These changes were not significantly different between groups (p > 0.05). Similar to control injections of glutamate, microinjection of glutamate 5 min following bicuculline also produced increases in MAP, HR, and LSNA. However, increases in MAP produced by glutamate injections at both 5 and 15 min after bicuculline were selectively enhanced in sedentary animals such that responses were significantly greater in sedentary versus physically active animals (p < 0.05 for main effects, interaction, and post-hoc tests). Glutamate-induced pressor responses at 30 and 45 min were similar to glutamate control responses suggesting recovery of responses and consistent with the time course of action of bicuculline in producing disinhibition of the RVLM (Miyawaki et al., 2002; Mueller, 2007). Interestingly, glutamate-induced increases in LSNA were also enhanced by bicuculline at 5 and 15 min in both groups (p < 0.05 for main effect and multiple comparisons within main effect) but were not significantly different between sedentary and physically active animals (p > 0.05 for main effect). At 30 and 45 min after bicuculline, increases in LSNA produced by glutamate were not significantly different than control glutamate responses (p > 0.05 for multiple comparisons within main effect) and were not different between groups, also suggesting full recovery of responses from bicuculline in a manner consistent with its time course of action. Lastly, bicuculline produced a slight but significant main effect on the small HR responses (p < 0.05 for main effect), but there was neither a main effect of sedentary condition nor interaction to allow testing for further differences (p > 0.05 for both).

To control for potential volume-, vehicle-, or time-related effects of bicuculline injections on responses to glutamate microinjections, we performed a series of experiments utilizing microinjections of the bicuculline vehicle (i.e., aCSF) before and after repeat microinjections of glutamate into the RVLM. As in Protocols 1 and 2, control responses to glutamate (30 nl, 10 mM) produced increases in MAP and LSNA and small changes in HR (<10 bpm) which were not different between sedentary and physically active animals. Similar to previous studies (Mueller, 2007), microinjections of aCSF (60 nl) had little or no effect on baseline MAP (Sedentary –Δ2 ± 2 mmHg; Physically Active Δ0 ± 1 mmHg), HR (Sedentary –Δ1 ± 3 bpm; Physically Active –Δ1 ± 1 bpm), or LSNA (Sedentary Δ1 ± 3%; Physically Active Δ0 ± 1%). Responses to repetitive microinjections of glutamate at 5, 15, and 30 min were also not significantly different than control glutamate microinjections for increases in MAP and LSNA, and small changes in HR (p > 0.05 for main effect of aCSF). Lastly, responses to repetitive microinjections were not significantly different between groups (p > 0.05 for main effect of physically active condition).

Microinjection sites marked with 2% Chicago sky blue dye (30 nl) were recovered in 16 sedentary and 17 physically active animals used in these studies. As in our previous studies (Mischel and Mueller, 2011), all microinjections sites were near the caudal pole of the facial nucleus and located ventral to the nucleus ambiguus, lateral to the pyramidal tract, and medial to the spinal trigeminal tract, which characterize the boundaries of the RVLM established by previous studies (Guyenet, 2006; Schreihofer and Sved, 2011).

Although previous studies have reported significant differences in baseline sympathetic nerve activity in various models of cardiovascular disease (Huber and Schreihofer, 2011; Mischel and Mueller, 2011; Silva and Schreihofer, 2011), comparisons of absolute levels of sympathetic nerve activity across different groups of animals warrants careful consideration (Guild et al., 2010; Burke et al., 2011). Since there were no differences in LSNA voltages between groups, we chose to express our data in terms of percent change to allow for an easier interpretation of the relative changes in sympathetic nerve activity. Furthermore, with our repetitive microinjection protocols we were able to examine nerve activity responses within animals over a time course that also demonstrated recovery of responses back to control values. Understanding that several important factors contribute to the voltage signal in a given recording (Guild et al., 2010; Burke et al., 2011) and group differences in the present study weren't significant, we have presented our nerve activity data in the more traditional and more easily interpreted format of percent change.

In order to complete our concentration response and repetitive microinjection protocols in the RVLM of rats while recording MAP, HR, and LSNA, we felt the use of anesthesia was necessary in these studies. The use of anesthesia allowed us to localize our RVLM microinjections more readily and successfully in a given animal since we were able to adjust the location of injection when necessary to achieve responses typically elicited from the RVLM. Furthermore, the use of triple-barrel glass micropipettes allowed us to inject different concentrations of glutamate or inject bicuculline or aCSF before and after multiple injections of glutamate without having to remove the micropipette from its position in the RVLM. Notwithstanding these advantages, we are fully aware that our use of anesthesia may have altered responses compared to those observed in conscious animals. Since we have been able to identify significant differences in our other studies using nearly identical preparations (Mueller and Hasser, 2006; Mueller, 2007; Mischel and Mueller, 2011), we are confident in the relative similarities and differences observed in the present study.

Given the time-consuming nature of our microinjections studies and the practicalities of generating the number of animals required for this study, we grouped experiments from animals that ran between 8 and 12 weeks. This can be considered a limitation of our study since it did not allow us to distinguish between potential differences between separate groups of animals that ran for 8 vs. 12 weeks. The time course by which functional alterations occur in brain regions important in cardiovascular regulation in response to physical (in)activity is worthy of consideration for future studies.

One caveat to our studies is that we used a volume and concentration of bicuculline based on previously published studies (Miyawaki et al., 2002; Moffitt et al., 2002; Horiuchi et al., 2004), including a study from our own laboratory in which we rigorously tested the effectiveness of GABA_A receptor blockade in the RVLM of physically active and sedentary rats (Mueller, 2007). Our expectation was that a similar volume (60 nl) and identical concentration (5 mM) of bicuculline used unilaterally would produce a similar level of blockade as to that used bilaterally in these previous studies. This line of thinking was reinforced by our experimental design in which we tested a higher volume of bicuculline (60 nl) against a smaller volume of glutamate (30 nl), both in the same micropipette, in order to maximize the chances that glutamate was only activating neurons in a region that had been affected by the antagonist. Nonetheless, the fact is that we did not perform experiments in the current study to confirm whether we achieved similar levels of blockade in both groups of animals. Consequently we cannot eliminate the possibility that GABA_A receptors were blocked to a lesser extent in the wheel running group and provide an explanation for the reduced effects of bicuculline on the blood pressure response to glutamate. We contend, however, that it is hard to reconcile the collective results of the current study with this possibility. For example, bicuculline produced similar effects on baseline MAP and SNA in both groups while producing differential effects on the glutamate-induced pressor response. Bicuculline also similarly enhanced the LSNA responses in both groups in a time course consistent with the known actions of bicuculline observed in previous studies (Miyawaki et al., 2002; Moffitt et al., 2002; Horiuchi et al., 2004; Mueller, 2007). Thus, it seems hard to conclude that these different effects could all occur because bicuculline produced less effective blockade in the wheel running group.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.